基于时频特征的光纤周界振动信号识别

发布时间：2018-10-26 09:06

【摘要】：在光纤周界安防系统中,蓄意入侵和环境噪声均能引起光纤传感器振动,在保证系统高灵敏度的前提下区分入侵和非入侵事件极为重要。为了有效识别各种光纤振动信号,本文依据入侵和环境噪声引起的光纤振动信号在时域上的短时特性以及复小波域各尺度上能量分布特征,提出了两级判别法识别光纤信号。第一级用时域特征,短时能量和短时平均过零率判断是否有振动发生;第二级用复小波提取光纤信号的能量分布特征,联合时域特征形成特征矢量,支持向量机(SVM)作为分类器识别是否为入侵信号及入侵类型。实验结果表明,此方法可以有效识别入侵信号和环境噪声引起的非入侵事件,提高了系统报警率,降低了误报率。
[Abstract]:In the optical fiber perimeter security system, intentional intrusion and environmental noise can cause vibration of optical fiber sensor. It is very important to distinguish intrusion and non-intrusion events under the premise of high sensitivity of the system. In order to identify all kinds of optical fiber vibration signals effectively, a two-stage discriminant method is proposed to identify optical fiber signals according to the short-time characteristics of the optical fiber vibration signals caused by intrusion and environmental noise and the characteristics of energy distribution at various scales in complex wavelet domain. In the first stage, the characteristics of time domain, short time energy and short time average zero crossing rate are used to judge whether there is vibration or not. In the second stage, the energy distribution of fiber signal is extracted by complex wavelet, and the feature vector is formed by combining time domain features. Support vector machine (SVM) is used as classifier to identify whether the signal is an intrusion signal or not. The experimental results show that this method can effectively identify the non-intrusion events caused by intrusion signal and environmental noise, improve the alarm rate of the system and reduce the false alarm rate.
【作者单位】：合肥工业大学电气与自动化工程学院;
【基金】：国家自然科学基金资助项目(51177034)
【分类号】：TN911.7;TH74

【参考文献】